185
stacked
71d

Real conversation:

Coworker: I'm trying to classify data based on X

Me: Mhh. Seems like a hard task, we don't have data to figure out X

Coworker: I know! That's why I thought about using machine learning!

Me: (Oh, boy)

Coworker: I'm working on training this ML model that will be able to classify based on X

Me: and what are the inputs for your training?

Coworker: The data classified based on X

Me: And where did you get that from?

Coworker: I don't have it! That will be the output of my ML model!

Me: But you just said that was the input!

Coworker: Yes

Me: Don't you see a contradiction here?

Coworker: Yes, it's a pretty complicated problem, that's why I'm stuck. Can you help me with that?

Me: (Looking at my watch) Sorry I'm late for a meeting. Catch up later, bye!

Comments
  • 59
  • 80
    ML: Magical Learning
  • 7
    Dear lord, I would burst into a laugh how the hell dou stayed alive
  • 5
    What the hell man 🤣
  • 48
    Tell him to use "more blockchain".
  • 9
    Deus ex machina works only in movies and books, and even then it sucks.
  • 15
    When you want to figure out X and create a machine learning model that needs X to figure out X you know it's time for a break
  • 1
    In all seriousness, can he not do something to reduce the dimensionality of the data and display it in an easier to see way. Using PCA and plotting the two highest values with the greatest variance would give a good representation of his data and maybe give insight on things he can use to classify it.
  • 3
    "You wot, mate?"
  • 1
    let input = X

    runML().then(output => {
    console.log(output); // also X 🌚
    });
  • 4
    @SoulSkrix no, we're still at an early stage for that. "X" here is whether certain events were considered important by our users or not, and the problem is that we don't have any system to collect the feedback from users.

    There's absolutely no way we can guess or visualize X unless we close the feedback loop. We can't even do the classification work manually ourselves because our users are the only domain experts, and sometimes even they lack the knowledge to judge the impact of an event.

    So, really, there's no technique or tool that can help here, besides A FUCKING SIMPLE FEEDBACK FORM for our users that I've been proposing for months now.
  • 3
    @stacked oh lmfao, I do not realise that was the issue (to be fair I likely didn't read the rant well enough).

    How can your coworker seriously expect to analyse data without any data. Ah.. wait... You can machine learning up some data!
  • 4
    What is deadlock?
  • 3
    I've had a PM once that also wanted from me to create model that detects objects on images that are like 8x8 px and often blury with at least 90% accuracy with training set of about 20 images. Lets just say he is no longer a PM.
  • 1
    Now you're thinking with temporal paradoxes!
  • 0
    W T F
  • 0
    WTF ML means for him....!!
Your Job Suck?
Get a Better Job
Add Comment